Hello,

I’m not sure that this is the right place for this question, but I hope my description is clear.

I’m trying to analyze the performance of a plano-convex aspheric lens coupling light into a SMF.  From the POP analysis in Zemax, the fiber coupling efficiency is farily high, ~80%.  We expect ~82% for coupling a top-hat beam into a single Gaussian mode, so this number is close to optimal for our case.

In the lab, we are currently getting ~40% coupling efficiency.  Upon doing some data analysis on our fiber, we measure a lower divergence than what the fiber was spec’d to.  Turns out, when splicing an endcap onto the tip of a fiber, the core expands to a higher MFD, resulting in a lower divergence.

I’m then assuming that the beam entering the fiber has to be mode-matched to the endcap, not the tip of the fiber.  So in Zemax, I have the last two surfaces set as the front surface of the endcap, propagated a known distance in a known material to the image surface, the fiber tip.  In the POP analysis, the fiber data has to then correspond to the MFD of the tip, not the endcap.  This model makes sense to me, yet I still get a coupling efficiency value that is much higher than what we get in the lab.

Does anybody have experience modelling fibers with endcaps to design a lens that will optimize coupling? I am trying to narrow down if this is just an issue of alignment in the lab, or pure design based on the way the fiber mode is modelled.

Thank you!